UNIVERSITA’ DEGLI STUDI DELL’INSUBRIA
DIPARTIMENTO DI MEDICINA CLINICA E SPERIMENTALE
DOTTORATO DI RICERCA IN
MEDICINA SPERIMENTALE E ONCOLOGIA
XXV ciclo
Coordinatore: Chiar.mo Prof. Antonio Toniolo
MOLECULAR ASPECT OF SMALL BOWEL ADENOCARCINOMA FOR NEW THERAPEUTIC
APPROACHES
Relatore: Chiar.mo Prof. FAUSTO SESSA
Correlatore: Chiar.mo Prof. LUCA MAZZUCCHELLI
Tesi di Dottorato di:
MILO FRATTINI
Anno Accademico 2012/2013
INDICE
ABSTRACT 4
INTRODUCTION 8
1.1 Cancers of the small intestine 9
1.2 Risk factors 11
1.3 Clinico-pathological data 13
1.4 Pathological features 14
1.5 Stage and grade 16
1.6 Prognosis 18
1.7 Clinical treatment 19
1.8 Molecular data 20
1.8.1 Microsatellite instability 20
1.8.2 APC-β-catenin pathway 20
1.8.3 KRAS 21
1.8.4 Allelic imbalance of Chromosome 18q 22
1.8.5 TP53 22
1.8.6 Molecular models of colorectal carcinogenesis 23
1.8.7 EGFR in colorectal cancer 24
1.8.8 EGFR downstream cascade 25
1.8.9 Targeting EGFR in colorectal cancer: anti-EGFR therapies 28 1.8.10 Molecular mechanism of response and resistance to EGFR
targeted monoclonal antibodies 29
AIM 33
PATIENTS AND METHODS 35
3.1 Patients 36
3.2 Patient treatment and clinical evaluation 36
3.3 Molecular analyses 37
3.4 Immunohistochemical analyses 37 3.5 Fluorescent in situ hybridization (FISH) 39
3.6 Mutational analysis by direct sequencing 43
3.7 MSI 44
3.8 Allelic imbalance of Chromosome 18q 45
3.9 Statistical analyses 47
RESULTS 48
4.1 Patients cohort 49
4.2 β-Catenin protein expression 51
4.3 Microsatellite instability analysis 52
4.4 KRAS gene mutations 53
4.5 Chromosome 18q analysis 54
4.6 TP53 gene mutations 55
4.7 EGFR gene status 57
4.8 BRAF gene mutations 59
4.9 PIK3CA gene mutations 60
4.10 PTEN protein expression 61
4.11 Correlations 63
4.12 Correlation of molecular data with clinical response to
EGFR-targeted therapy 69
DISCUSSION 71
REFERENCES 81
ACKNOWLEDGMENTS 91
ABSTRACT
Small Bowel Adenocarcinoma (SBA) is a rare and aggressive neoplastic disease.
Several aspects of SBA carcinogenesis still need to be elucidated, but risk factors and histomorphological similarities seem to indicate that SBA can follow a carcinogenetic development similar to that proposed for colorectal cancer. At molecular level, at odds with adenocarcinoma arising in the large intestine, very few, and fragmented, information is available for SBA. In general, it has been suggested that the two models of colorectal carcinogenesis can be valid also for SBA. For that reason, chemotherapies set up for the cancers of the large intestine have been applied also for SBA. Therefore, since recent studies led to the introduction of EGFR-targeted therapies in colorectal cancer, the treatment with anti-EGFR drugs can be proposed also for SBA patients. In particular, in colorectal cancer patients it has been demonstrated that KRAS mutations are correlated with the absence of efficacy of EGFR-targeted therapies, and it has been proposed by few studies to investigate additional markers of the EGFR pathway (EGFR gene copy number, BRAF and PIK3CA mutations, as well as PTEN protein expression), in order to increase the predictive power of the efficacy of anti-EGFR drugs. Information regarding these markers in SBA is quite completely missing.
Primary aim of the present work was the evaluation, in the same cohort, of all the alterations involved in colorectal carcinogenesis, in order to shed light more deeply about the molecular similarity between SBA and colorectal cancer. Second aim of the present work was to investigate in the same cohort of SBA the aforementioned markers involved in the EGFR pathway, in order to verify if the pattern of these alterations could justify the possible introduction of these therapies also in patients affected by SBA.
To do this, for the first aim we investigated β-catenin protein expression by immunohistochemistry (IHC), and KRAS and TP53 mutations by direct sequencing, as well as microsatellite instability (MSI) and allelic imbalance of Chromosome 18q MSI by fragment analysis on genomic DNA extracted from formalin-fixed paraffin- embedded tissue sections. For the second aim we investigated EGFR gene status by
Fluorescent in situ hybridization (FISH), BRAF and PIK3CA mutational status by direct sequencing, as well as PTEN protein expression by IHC, in the same cohort of SBA.
We recruited 40 SBA cases from the Institute of Pathology of Locarno (Canton Tessin, Southern Switzerland) and from three institutions of Northern Italy.
First aim. β-catenin overexpression was observed in 23.6% at nuclear level and in additional 47.3% of cases only at cytoplasmic level, MSI was found in 23.6% of cases, KRAS mutations in 43.6% of cases, TP53 mutations in 29% of cases and allelic imbalance of Chromosome 18q in 75% of cases. All the percentages of alterations and the types of mutations are in line with those identified in the analysis of colorectal cancer patients. Therefore, by the analysis of all these markers in a same cohort, we can confirm that SBA shares the carcinogenetic development with colorectal cancer also at molecular level.
Second aim. We identified a copy number gain of EGFR gene in 57.5% of cases, BRAF and PIK3CA mutations in 2.5% and 10.5% of cases respectively, and PTEN loss of expression in 25.6% of cases. Also for the EGFR pathway analysis, percentages of alterations and types of mutations found in SBA are in line with colorectal cancer, even if we did not detect the classical V600E change in the BRAF gene (where, on the contrary, we found a rare mutation, the G596R change). Taking into account the molecular algorithm proposed for the administration of EGFR-targeted therapies in colorectal cancer patients, if we look only at KRAS mutations, we can propose the administration of EGFR-targeted therapies to about 60% of patients (i.e.: KRAS wild- type cases), and to 23% of cases if we base our evaluation on the whole EGFR pathway (i.e.: cases showing, at the same time, EGFR copy number gain, KRAS, BRAF and PIK3CA wild-type sequences, and PTEN normal expression). The treatment with anti-EGFR therapies of a SBA patient of our cohort who developed a metastatic lesion confirmed the relevance of the molecular characterization of the tumor to predict the response to these therapies.
In conclusion, our analyses of SBA confirm the feeling that the mechanisms of carcinogenesis of such disease are superimposable with those proposed for colorectal cancer. Therefore, the hypothesis that therapeutic protocols valid for the large intestine can be applied also to SBA patients is supported. As a consequence, the targeted therapies recently introduced in colorectal cancer can be proposed for SBA patients, pending tumor molecular characterization as demonstrated by our case report.
INTRODUCTION
1.1 Cancers of the small intestine
The small intestine represents the longest part of the digestive tract, making up 75%
of the length (about 6 m long and 4 times as long as the large intestine) and 90% of the absorptive surface area of the gastrointestinal tract (Figure 1.1). It has three sections: duodenum, jejunum and ileum (Figure 1.2). Malignant tumors of the small intestine are rare all over the world (Hamilton & Aaltonen, 2000), especially with respect to tumors arising in the other portions of the gastrointestinal tract, with a global incidence of less then 1.0 per 100,000 population (Curado et al, 2009). Cancers of the small intestine, including adenocarcinoma, carcinoid, lymphoma and sarcoma (the two latter are rarer than adenocarcinoma and carcinoid) account for only 0.42%
of total cancer cases and 2.3% of cancers of digestive system in the United States (Jemal et al, 2009). Mortality of the cancer is even lower, accounting for only 0.2% of the total cancer deaths in the United States (Jemal et al, 2009).
Figure 1.1: Gastrointestinal tract.
Approximately 30%-40% of the cancers observed in the small bowel are adenocarcinomas (SBA), a percentage much lower than the proportion in the colon where the overwhelming majority is adenocarcinomas (Haselkorn et al, 2005;
Bilimoria et al, 2009; Schottenfeld et al, 2009). The incidence rates for SBA are 0.5-1.5 per 100,000 in men and 0.2-1.0 per 100,000 in women. High incidence rates are observed in black people in several regions of the United States: more than 2 per 100,000 men and about 1.25 per 100,000 women in the regions included in the Surveillance, Epidemiology and End Results (SEER) program. High rates are also observed in Hawaii and in New Zealand. On the contrary, the lower incidence rates can be found in India, Romania and in other countries of Eastern Europe (Neugut et al, 1998; Negri et al, 1999). From 1975 to 2000, it seems that the incidence rates are increased of 50% in several countries, but not in the United States (De Launoit al, 2005). More than 50% of SBA arise in the duodenum, while ileum represents the rarest affected region, with the exception of patients affected by Crohn’s disease.
Mean age of SBA occurrence is between 55 and 65 years, even if a few cases have been described in younger patients, especially in cases belonging to families with a inherited history of colorectal cancer, or with Cronh’s disease (Lashner et al, 1992).
Figure 1.2: Small intestine.
1.2 Risk factors
The reason for the much lower incidence of small intestinal cancer than of colorectal cancer is largely unknown but has been hypothesized to be related to several mechanisms. The much quicker transit time of food in the small intestine than in the large intestine (because peristaltic ring contractions in the small intestine occur with greater frequency than in the colon) may result in shorter time of exposure of its mucosa to carcinogens. The small intestine has much lower bacterial load, thus has decreased concentration of potential carcinogens from bile acid breakdown (Arber et al, 1997). Studies also demonstrate that the small intestine generates less endogenous reactive oxidative species than the colon does, which may lead it to handle oxidative stress more effectively than the colon thus resulting in less oxidative damage during times of exposure to oxidant stress (Sanders et al, 2004).
Inflammatory bowel disease includes Crohn’s disease and ulcerative colitis, two clinically related but histologically distinct diseases. Crohn’s disease is a recognized risk factor for SBA, with relative risks reported as high as 60 (Neugut et al, 1998; Pan et al, 2011). A meta-analysis showed a relative risk of 33.2 (95% CI: 15.9-60.9) for SBA in patients with Crohn’s disease (Canavan et al, 2006). Extended duration of the disease, distal jejunal and ileal location, male sex, small bowel bypass loops, chronic fistulous disease, young age of diagnosis and occupational hazards or exposure to halogenated aromatic compounds with aliphatic amines, asbestos and solvents are suggested to be associated with an increased risk of SBA in patients with Crohn’s disease (De Launoit al, 2005; Feldstein et al, 2008). Ulcerative colitis has been shown to be associated with an increased risk of colorectal cancer, hepatobiliary cancer, nonmelanoma skin cancer and leukemia (Winther et al, 2004; Hemminki et al, 2008).
However, it is unclear whether patients with ulcerative colitis have an increased risk also for SBA (Bernstein et al, 2001; Hemminki et al, 2008).
Celiac disease is an inflammatory small intestinal disorder characterized by the inability of the small intestine to deal with the gluten fractions of cereals such as wheat, barley and rye; its prevalence is nearly 1% of general population (Pan et al,
2011). The risk of SBA in patients with celiac disease is increased many-fold as compared with the risk in the general population (Green and Cellier, 2007) with reported relative risks between 60 and 80 (Green et al, 2003). SBA is most often located in the jejunum and is more likely to develop as an adenoma-carcinoma sequence than as dysplasia in flat mucosa (Green and Rampertab, 2004).
Although the prevalence of adenomas in the small intestine is much lower than their prevalence in the colon, it is suggested that the adenoma-carcinoma sequence is as significant in the small intestine as in the large intestine (Sellner, 1990). As in the colon, adenoma in the small intestine appears to be a precursor of adenocarcinoma (Gill et al, 2001). A large fraction of villous adenomas of the small intestine has been shown to progress to malignancy (Bjork et al, 1990).
Familial adenomatous polyposis (FAP) is an autosomal dominant genetic disorder caused by mutations of the APC gene on the long arm of chromosome 5 (Groden et al, 1991). Most patients diagnosed with FAP have multiple adenomas in the small bowel, usually in the duodenum (Bertoni et al, 1996) and these patients are at increased risk of SBA, especially duodenal cancer (Lepisto et al, 2009). The prevalence of duodenal adenomatosis in FAP patients are 50%-90% and 3%-5% of these patients develop duodenal cancer (Kadmon et al, 2001).
The demonstration of a geographical correlation between rates of SBA and colorectal cancer (Haselkorn et al, 2005) suggests a common aetiology. Various studies have shown that the risk of SBA following primary colorectal cancer were elevated; in addition, in those diagnosed with primary SBA, there was a 4 to 5-fold risk of developing colorectal cancer (Murray et al, 2004; Scelo et al, 2006; Lagarde et al, 2009). These studies suggest etiological similarities between adenocarcinomas of the small intestine and of the colon-rectum but, to date, potential common carcinogenic agents have not been elucidated in analytic epidemiological studies. Dietary factors have been suggested to be related to the risk of SBA. A study of SBA mortality and food data by WHO showed correlations with daily consumption of animal fat and animal protein (Lowenfels and Sonni, 1977). A case-control study of 430 SBA cases
and 921 controls observed two-to three-fold increases in SBA risk with frequent intake of red meat and salt-cured/smoked foods but no association with alcohol consumption (Chow et la, 1993). Another case-control study of 36 cases with SBA and 998 population controls also reported a significant increase in risk associated with frequent intake of foods rich in heterocyclic aromatic amines (based on the combined intake of fried bacon and ham, barbecued and/or smoked meat and smoked fish) in males only and with total sugar intake (Wu et al, 1997). A hospital- based case-control study (Negri et al, 1999) found an increased risk of SBA among the highest consumers of red meat and of refined carbohydrates, while a decreased risk was associated with consumption of fish and vegetables.
1.3 Clinico-pathological data
Symptoms predicting the insurgence of SBA depend on the dimension and on the site of the lesion. The majority of these tumors are characterized by aspecific symptoms, able to be perfectly understood only when the disease is at advanced stage. This unlike factor leads to a late diagnosis, and as a consequence the prognosis is often severe. A late diagnosis of 6-8 moths is a common event for two additional reasons:
- SBA can be hardly detected by the endoscope, especially when it occurs in the duodenum;
- X-ray exam is not the preferred methods to identify these lesions (Dabaja et al, 2004).
SBA in the ileum and jejunum is characterized by abdominal pain, nausea, vomiting, weight loss (Schottenfeld et al, 2009). Duodenal SBA has different clinical features, especially due to the fact that lumen is larger than in the ileum or jejunum (Longacre et al, 1990).
1.4 Pathological features
Duodenal SBA is typically polypoid with a necrotic central area (Figure 1.3). Very often an adenomatous component is still present. SBA arising near Papilla of Vater is usually smaller than that arising in the other regions of the small bowel, and appears to be a sort of nodule in the duodenum wall(Longacre et al, 1990).
a b
a b
Figure 1.3: Duodenal SBA with the typical polypoid aspect.
SBA of the ileum or jejunum is usually identified at higher stage with respect to duodenal SBA, and therefore the typical appearance is represented by an infiltrative and ulcerated mass protruding in intestinal lumen. In the majority of cases sierosa is infiltrated(Bridge et al, 1975) (Figure 1.4).
SBA occurring in patients affect by Crohn’s disease is difficult to be detected because of the presence of deep ulcerative lesions, generally undistinguishable from an inflammatory disease(Horton et al, 1994).
a b
a b
Figure 1.4: SBA with intestinal wall infiltration. a: lumen; b: sierosa. Arrows indicate intestinal wall infiltration
At histological level, SBA (Figure 1.5) is similar to adenocarcinoma of the colon- rectum.
a b
a b
Figure 1.5: Two examples of microscopical appearance of SBA.
1.5 Stage and grade
Classification criteria subdivide SBA in well, moderately and poorly differentiated, thus identifying tumor grade. Well differentiated tumors (G1) (Figure 1.7) are characterized by glandular structures in more than 95% of tumor mass, moderately differentiated cases (G2) between 50 and 95% (Figure 1.8), poorly differentiated tumors (G3) by less than 50% (Figure 1.9). Undifferentiated tumors have only less than 5% of glandular structures (Schlemper et al, 2000). Mucinous adenocarcinoma are considered G3 cases.
Figure 1.7: Well differentiated SBA (G1).
Figure 1.8: Moderately differentiated SBA (G2).
Figure 1.9: Poorly differentiated SBA (G3).
In the presence of tumor heterogeneity, that means that different components coexist in the same tumor, the case is classified on the basis of the higher tumor grade component. However, if poorly differentiated cells are present only at the invasive margin, this feature is not sufficient to classify the tumor as G3.
Tumor staging is similar to that applied for colorectal cancers. UICC classification is based on tumor size and infiltration (T), regional lymph nodal status (N) and presence of distant metastasis (M). This system, also named as TNM, was lastly modified in 2010 and is generally accepted.
The tumoral lesion is considered as adenocarcinoma also in absence of muscolaris mucosa invasion, at odds with criteria valid for colorectal cancer, because in the small bowel lymphatic vessels are present very close to the epithelium, while in the large intestine they are localized in the muscolaris mucosa. Therefore, the possibility of tumoral cells to invade regional lymph nodes is higher in small bowel than in colon-rectum.
1.6 Prognosis
According to the data from US SEER for the period 1992 to 2005, the median 5-years relative survival was 28.0% for SBA in general and 32.5% for patients who underwent resection. This difference emphasizes the benefit of tumor resection on patients’ follow-up. Overall, only a minor fraction of patients can survive to SBA, and the cause of the severity of such a neoplastic disease can be principally ascribed to the delay of disease discovery (Dabaja et al, 2004). Although other cancer sites have demonstrated higher long-term survival rates due to novel adjuvant therapies over the last two decades, the US data from 1985 to 2000 showed no significant change in long-term survival rates for SBA (Bilimoria et al, 2009). Another study observed an improvement in survival rates in England, Wales and Scotland over the time period of 1975 to 2002 but the changes were not statistically significant because of the small number of patients (Shack et al, 2006). A Swedish study found 5-years survival rates of 39% for duodenal adenocarcinoma and 46% for jejuno-ileal adenocarcinoma (Zare et al, 1996). Earlier tumor stages at diagnosis (stage I and II), small tumor size and curative resection have been identified as factors for favorable overall survival, whereas poorly differentiated tumors, lymph node involvement or metastasis and distant metastases as factors predicting poor prognosis (Wu et al,
2006; Hamilton et al, 2009). However, the involvement of regional lymph nodes as prognostic factor is still debated, as other studies reported absence of worse survival for advanced cases with respect to patients with a localized disease (Willet at al, 1993;
Agrawal et al, 2007).
Primary relapse events in SBA patients are the loco-regional recurrence or liver metastasis. More rarely, peritoneal carcinomatosis can occur. Usually, relapse develops in the first 2 years after tumor resection (Pilati et al, 2001; Bilimoria et al, 2009).
1.7 Clinical treatment
Primary treatment is surgery, adopted whenever possible. Duodenal lesions are resected on the basis of Whipple’s indications. However, due to the late diagnosis, very often tumors cannot be resected. A good estimation is that only 50-60% of patients with SBA can be radically resected(Bilimoria et al, 2009).
Tumor relapse, due either to local recurrence or to the presence of distant metastatic lesions, therefore represents the main cause of death for SBA patients. As a consequence, major efforts have been done for the identification of the best chemotherapy. At the moment, however, a specific and standard protocol for the treatment of SBA has not been established yet. In general, oncologists prefer to apply the protocols set-up for cancers arising in other district of the gastrointestinal tract, such as for gastric and, especially, colorectal cancer.
The use of cisplatin turned out to be quite inefficacious (Ono et al, 2008; Suenaga et al, 2009). On the contrary, treatments with good results in SBA patients include oxaliplatin or irinotecan in conjunction with 5-fluorouracil (Overman et al, 2008). In addition, it has been demonstrated that the combination of 5-fluoruracil, doxorubicin and mytomicin C is active and well tolerated in advanced cases (Gibson et al, 2005).
However, all these combinations did not lead to a substantial benefit for SBA patients. Therefore, it is of particular interest the introduction of new therapeutic approaches for such a disease. In this category, we can easily include targeted
therapies, which turned out to be effective in colorectal and gastric cancer. But the possibility to introduce these therapies requires an extensive knowledge of the disease at molecular level, at least for the targets of these new compounds.
1.8 Molecular data
The data concerning the molecular characterization of SBA are few and fragmented, especially due to the low frequency rate of such a disease. In the majority of cases, only few patients are included in the analyses, and usually only few markers are investigated in the same cohort of cases. The most studied markers are those involved in colorectal carcinogenesis, i.e.: microsatellite instability, APC, KRAS, allelic imbalance of Chromosome 18q, and TP53.
1.8.1 Microsatellite instability
Since SBA can develop in Hereditary Non-Polyposis Colorectal Cancer (HNPCC) families, a syndrome characterized by a not functioning DNA mismatch repair system which leads to the accumulation of errors in all over the genotype and especially in small regions named microsatellites, the status of microsatellite instability (MSI) has been investigated. The majority of studies agrees in considering around 10% the fraction of SBA with MSI, a percentage in keeping with what observed in colorectal carcinogenesis (Keller et al, 1995; Svrcek et al, 2003). Another study found a higher percentage of MSI (35%) (Overman et al, 2010).
1.8.2 APC-β-catenin pathway
Since SBA can occur also in FAP patients, the occurrence of APC mutations has been studied. The APC gene is a tumor suppressor gene encoding for a large multidomain protein that plays a relevant role in the wnt-signalling pathway and in intercellular adhesion. In the normal cells, APC is able to form a multiprotein complex with GSK- 3β and axin. This complex binds to β-catenin, which in turn is phosphorylated by GSK-3β and subsequently degraded by the proteasome pathway (see Figure 1.10). In
tumoral cells, when APC (as well as β-catenin or axin) is mutated, the multiprotein complex cannot be formed and, therefore, β-catenin accumulates into the cytoplasm
and translocates into the nucleus, where activates Tcf factor, causing transcription of target genes (involved in different cellular processes), such as c-myc. {4991}. In the literature it seems that APC mutations are a rare event in SBA carcinogenesis, at odds with colorectal cancer, where APC is mutated in about 80% of sporadic cases (Abrahams et al, 2002; Svrcek et al, 2003).
Figure 1.10: APC/β-catenin pathway and myc overexpression in normal (left) and tumoral (right) cells.
1.8.3 KRAS
KRAS gene encodes for a membrane-bound 21 kd protein involved in G protein- mediated signal transduction. KRAS protein can acquire transforming potential secondary to a point mutation in hot spot codons, primarily codons 12 and 13, which
SBA the percentage of KRAS mutation occurrence is superimposable with that found in colorectal cancer (30-40%) (Sutter et al, 1996; Younes et al, 1997). In particular, it has been proposed that KRAS mutations are limited to the tumors arising in the duodenum, since have not been observed in tumors of ileum or jejunum (Younes et al, 1997).
1.8.4 Allelic imbalance of Chromosome 18q
An old study reported the loss of the long arm of Chromosome 18q in SBA (Hahn et al, 1996). Mapping to this chromosomal bands are three tumor suppressor genes mainly involved also in colorectal cancer: Deleted in Colon Cancer (DCC), SMAD4 (previously named DPC4, as Deleted in Pancreatic Cancer locus 4) and SMAD2. The DCC protein plays a relevant role in the regulation of cell-cell and cell-matrix adhesion, while SMAD proteins are involved in cell proliferation and apoptosis. In SBA, the loss of this region can be observed in a wide range of cases, from 15 to 80%
of analyzed patients (Bläker et al, 2002; Bläker et al, 2004), similarly to what observed in colorectal cancer.
1.8.5 TP53
The TP53 gene encodes a 393-aminoacid nuclear phosphoprotein, which negatively controls the cell cycle through transcriptional activation of WAF1/CIP1 gene and through bcl-2 and Bax binding in response to a variety of stress signals including DNA damage as well as hypoxia, radiation exposure, drug exposure. TP53 mutations, especially occurring in the DNA binding domain (encoded by exons 5-8), represent the main mechanism of TP53 inactivation in cancer. Studies investigating TP53 mutations in SBA reported their detection in 20-53% of cases, a percentage in keeping with colorectal carcinogenesis (Lane, 1994; Abrahams et al, 2002; De Launoit et al, 2004).
1.8.6 Molecular models of colorectal carcinogenesis
Intensive screening for genetic alterations led to the identification of two major types of colorectal cancer, that are distinct by their carcinogenic process (Figure 1.11). One is characterized by normal caryotype, normal DNA index (Houlston et al, 2001) and genetic instability at microsatellite loci and is called MSI-positive cancer (Ilyas et al, 1999). The second one, valid for more than 90% of sporadic CRC and firstly proposed by Vogelstein’s group (Fearon and Vogelstein, 1990), suggests that APC (or, better, the APC-β-catenin pathway) represents the initial mutational event that determines
hyperplastic proliferation and then early adenoma formation. The stage of late adenoma is achieved with KRAS mutation. Loss of tumor suppressor genes at chromosome 18q and mutations in TP53 gene lead to carcinoma in situ (Laurent-Puig et al, 1999) and then to metastasis.
Since in SBA the percentages of alterations of these markers are similar to those observed in colorectal cancer, it is generally accepted that the two models proposed for colorectal cancers are valid also for SBA and, therefore, that SBA carcinogenesis mirrors that of the colon-rectum.
Figure 1.11: Schematic models of colorectal carcinogenesis as proposed by Vogelstein and colleagues
However, these markers have not been introduced in clinical practice in the management of colorectal cancer patients, where, on the contrary, recent evidence pointed out to the relevance of the EGFR pathway.
1.8.7 EGFR in colorectal cancer
The EGFR signaling pathway is thought to play a pivotal role in tumor growth and progression of various cancers, including CRC. The EGFR gene encodes for a 170 kDa transmembrane receptor with intrinsic tyrosine kinase activity belonging to the ErbB family of receptor TKs [that includes ErbB1 (EGFR or HER1), ErbB2 (HER2), ErbB3, and ErbB4] (Figure 1.12).
EGFR binds to, and then is activated, by several ligands, leading to receptor dimerization, which in turn is able to transmitting the mitogenic signaling through
several pathways into the nucleus by regulating several transcription factors. EGFR is involved in the control of the expression of genes relevant for inhibition of apoptosis and for tumor cell proliferation and survival, migration, adhesion and angiogenesis (Woodburn 1999; Arteaga, 2001; Talapatra et al, 2001; Venook, 2005).
Figure 1.12. Structure of the EGFR protein (A), activation (B) and dimerization by ligand binding (C) (Mitsudomi and Yatabe, 2009).
In colorectal cancer, the main mechanism of EGFR deregulation is represented by protein overexpression following gene amplification or at least an increase of the gene copy number.
1.8.8 EGFR downstream cascade
The two main pathways activated by EGFR are the RAS-RAF-MAP kinase pathway, mainly involved in cell proliferation, and the PI3K-PTEN-Akt pathway, mainly involved in cell survival and escaping from apoptosis. In these two pathways, some members are more deeply involved in colorectal cancer development: KRAS, BRAF, PI3K and PTEN (Figure 1.13).
Figure 1.13. Schematic representation of the two main pathways of EGFR downstream cascade.
KRAS. See paragraph 1.8.3.
BRAF. BRAF gene encodes for a RAS effector belonging to the RAF family of Ser-Thr kinase proteins. BRAF gene product is recruited to the plasma membrane upon binding to RAS-GTP, and represents a key point in the signal transduction through the MAP kinase pathway. The typical BRAF alteration, leading in turn in the constitutive activation of BRAF itself and of the MAP kinase pathway, is represented by point mutations, occurring in the vast majority of cases in colorectal cancer (>90%) at codon 600 (with the typical V600E change) (Davies et al, 2002). In colorectal cancer, BRAF mutations are frequently found in sporadic cases characterized by MSI, and are mutually exclusive with KRAS mutations (Rajagopalan et al, 2002). At the moment, no data concerning the involvement of BRAF in SBA cancerogenesis have been published.
PI3K. Phosphatidylinositol 3-kinases (PI3Ks) belong to the lipid kinases family that regulates the signal transduction (Vivanco et al, 2002). Activation of PI3Ks results in the production of the second messenger phosphatidylinositol (PI) 3,4,5 trisphosphate (PIP3) from PI 4,5 bisphosphate (PIP2). PIP3, through AKT activation, drives various
downstream pathways involved in the regulation of several cellular functions including cellular growth, transformation, adhesion, apoptosis, survival and motility (Yuan et al, 2008).
Only PI3K proteins that contain the catalytic subunit p110α and its associated regulatory subunit p85 (that belongs to the class IA protein) are involved in tumorigenesis (Samuels et al, 2004). The p110α subunit is encoded by PIK3CA, which in tumors is frequently iperactivated following point mutations in hot spot codons located in exons 9 and 20.
In colorectal cancer, PIK3CA mutations occur in about 10-30% of cases (Samuels et al, 2004), while their occurrence in SBA has never been investigated.
PTEN. PTEN is a tumor suppressor gene that encodes for a 403-amino acid protein that possesses both lipid and protein phosphatase activities. Its typical function consists of dephosphorylation of PIP3 and PIP2, thus preventing AKT phosphorylation, and maintaining it in its inactive form, thus counteracting the role of PI3K proteins. PTEN is therefore involved in inhibition of cell cycle progression, induction of cell death, modulation of arrest signal and stimulation of angiogenesis by influencing vascular endothelial growth factor activity and suppression of destabilization of hypoxia-inducible factor-1 (Sansal et al, 2004).
In colorectal cancer PTEN is altered through mixed genetic/epigenetic mechanisms (intragenic mutation/epigenetic or 10q23 loss of heterozigosity (LOH)/epigenetic), which lead to the biallelic inactivation of the protein in 20-30% of cases. In addition to PTEN LOH and mutations, PTEN promoter hypermethylation is a frequent event in MSI sporadic colorectal cancer and may constitute an important epigenetic mechanism of PTEN inactivation in this setting (Goel et al, 2004). All these alterations lead to the loss of PTEN protein expression and can be altogether analysed by protein expression analyses methods such as western blot (on cells or fresh/frozen tissues) or immunohistochemistry (IHC) on archival fixed tissues. No data have been published about PTEN role in SBA.
Overall, mutations in KRAS, BRAF and PIK3CA genes, as well as the loss of PTEN protein function result in continuous activation of EGFR downstream pathways, regardless of whether the EGFR is activated or pharmacologically blocked.
1.8.9 Targeting EGFR in colorectal cancer: anti-EGFR monoclonal antibodies
Given the important role of EGFR and its downstream pathways in tumorigenesis and disease progression, this receptor has become a relevant and promising target for anti-cancer therapies. In vitro and in vivo studies showed that blocking EGFR and downstream signaling may lead to carcinoma cell growth inhibition, resulting in potential benefits for cancer patients. In colorectal cancer, monoclonal antibodies (MoAb) targeting EGFR, namely cetuximab and panitumumab, have been developed and introduced in clinical practice (Rocha-Lima et al, 2007; Ciardiello and Tortora, 2008). Cetuximab, a human–mouse chimeric IgG1 MoAb, was the first EGFR- targeted agent approved for the treatment of colorectal cancer. Panitumumab, a fully human IgG2 MoAb was recently approved in the US and Europe as third-line treatment of metastatic colorectal cancer (Jonker et al, 2007; Amado et al, 2008;
Ciardiello and Tortora, 2008). Cetuximab and panitumumab bind to the extracellular domain of EGFR when it is in the inactive configuration, compete for receptor binding by occluding the ligand-binding region, and thereby block ligand-induced EGFR activation, inducing its internalization and degradation (Ciardiello and Tortora, 2008). Consequently, they block the activation of the EGFR mitogenic signal transduction pathways, and they inhibit therefore tumor cell proliferation, angiogenesis, invasion and metastatic spread by inducing apoptosis. Additionally, anti-EGFR MoAbs, particularly those of the IgG1 subclass, may recruit host immune functions to attack the targeted cancer cell. These functions include antibody- dependent cellular cytotoxicity and, to a lesser extent, complement-mediated cytotoxicity (Kimura et al, 2007; Kurai et al, 2007). Anti-EGFR MoAbs recognize EGFR exclusively and are therefore highly selective for this receptor. Most studies
have been made using cetuximab, but same results are valid also for panitumumab.
The ability of cetuximab for blocking the EGFR pathway is supported by preclinical and clinical studies. At preclinical level, it has been demonstrated that cetuximab alone primarily shows cytostatic activity, whereas its combination with other chemotherapeutic agents (such as platinum derived compounds and irinotecan) leads to synergistic antitumoral activity (Fan et al, 1993; Ciardiello et al, 1999; Baselga et al, 2000; Prewett et al, 2003). At clinical level, two phase II trials demonstrated that patients with advanced CRC had a response rate of 11% when cetuximab is administered as single agent therapy, and 23% when combined with irinotecan (Saltz et al, 2004; Cunningham et al, 2004; Chung et al, 2005). Both antibodies have been shown to reduce the risk of tumor progression and to improve overall survival (OS), progression-free survival (PFS) and quality of life of patients with refractory metastatic colorectal cancer (Saltz et al, 2004; Cunningham et al, 2004; Jonker et al, 2007).
1.8.10 Molecular mechanism of response and resistance to EGFR targeted monoclonal antibodies
From early clinical studies conducted mainly in heavily pretreated chemotherapy- refractory patients and also in chemotherapy-naïve patients with advanced colorectal cancer, it became clear that only 10% to 20% of patients with mCRC clinically benefited from anti-EGFR MoAbs (Cunningham et al, 2004; Saltz et al, 2004; Chung et al, 2005). This evidence, together with the side effects and the higher costs of MoAb therapies as compared with standard chemotherapy regimens, underlined the importance of studying the molecular mechanisms of primary resistance to cetuximab or panitumumab.
Initially, it was hypothesized that EGFR targeted agents would be most effective in those tumors overexpressing the EGFR protein (Vogel et al, 2002; Arteaga, 2003).
Preclinical studies demonstrated that anti-EGFR agents may have little activity when the level of EGFR expression is below a threshold level (Venook et al, 2005).
Consequently, cetuximab was indicated only for the treatment of patients who have tumors that demonstrated EGFR expression (Cunningham et al, 2004; Saltz et al, 2004; Chung et al, 2005). However, the immunohistochemical expression of EGFR protein turned out to be not a reliable tool for the identification of patients to be treated with EGFR MoAb, due to several reasons: the type of fixative used, the storage time of unstained tissue sections, the type of primary antibody used and the methods of IHC evaluation might generate conflicting data in the EGFR assessment (Atkins et al, 2004; Langner et al, 2004; Kersting et al, 2006). The same was observed with panitumumab-treated patients (Gibson et al, 2006; Siena et al, 2007; Van et al, 2007).
It was next investigated whether alterations of EGFR at gene status level might be predictive of anti-EGFR MoAbs efficacy. Several studies showed that EGFR gene copy number as detected by Fluorescence In Situ Hybridization (FISH) rather than EGFR protein expression evaluated by IHC, might better predict cetuximab response in advanced colorectal cancer (Moroni et al, 2005; Lièvre et al, 2006, Frattini et al, 2007; Sartore-Bianchi et al, 2007; Cappuzzo et al, 2008). However, a recent contribution demonstrated that FISH seems to be not a reproducible method to evaluate the EGFR gene status (Sartore-Bianchi et al, 2012). Therefore, at the moment, EGFR gene status is not evaluated as a marker useful for the prediction of anti-EGFR therapies efficacy.
On the contrary, there is a vast consensus for the use of KRAS mutations in clinical setting. This type of alteration, leading to the constitutive activation of EGFR downstream pathways, has been clearly demonstrated to be a negative predictor of cetuximab/panitumumab efficacy (Siena and Bardelli, 2010). These data have been included in the Food and Drug Administration (FDA) of the United States, and in the European Medicine Agency (EMA) guidelines. Therefore, at the moment, only patients with KRAS wild-type colorectal cancer can be treated with EGFR-targeted therapies.
Since BRAF, PIK3CA and PTEN play a superimposable role with respect to KRAS in the activation of EGFR downstream pathways, it was proposed that also BRAF and PIK3CA mutations, as well as the loss of PTEN protein expression might identify patients who are resistant to anti-EGFR therapies. Although promising reports opened interesting perspectives confirming this assumption, thus hypothesizing the possibility to identify more than 70% of resistant patients when all the abovementioned markers were investigated simultaneously (Frattini et al, 2007; Di Nicolantonio et al, 2008; Sartore-Bianchi A et al., 2009) (Figure 1.14), other and more recent contributions did not confirmed totally these preliminary data. In particular, it seems that BRAF mutations play a prognostic rather than a predictive role in patients treated with EGFR-targeted therapies, that not all PIK3CA mutations have the same role (no predictive effect for exon 9 mutations, at odds with those arising in exon 20) and that the loss of PTEN expression should be investigated in the metastatic lesion rather than in the corresponding primary tumor (Loupakis F. et al., 2009; Prenen et al, 2009; De Roock et al, 2010; Custodio et al, 2013). For all these considerations, BRAF, PIK3CA and PTEN deragulations are not investigated before the administration of EGFR-targeted therapies.
Figure 1.14. Proposed algorithm for a better prediction of EGFR-targeted therapies in metastatic colorectal cancer patients
Wild-type BRAF and
Wild-type PIK3CA and
Normal PTEN High probability
of clinical benefit Very low or no probability
of clinical benefit
Evaluation of KRAS status
Evaluation of BRAF, PIK3CA, and PTEN status Mutated BRAF
or
Mutated PIK3CA or
Loss of PTEN
Mutated KRAS Wild-type KRAS
AIM
Small bowel adenocarcinoma (SBA) is a rare disease presenting high similarity (for histology, epidemiology and risk factors) with colorectal adenocarcinoma. At molecular level, however, little is known for SBA, although it has been proposed that SBA shares the carcinogenetic development with colorectal cancer. For that reason, chemotherapies set up for the cancers of the large intestine have been applied also for SBA.
Primary aim of the present work was the evaluation, in the same cohort, of all the alterations involved in colorectal carcinogenesis, in order to shed light more deeply about the similarity at molecular level between SBA and colorectal cancer. To do this, we will evaluate MSI, β-catenin protein expression, KRAS and TP53 mutations, as well as allelic imbalance of Chromosome 18q.
Recent studies led to the introduction of EGFR-targeted therapies in colorectal cancer. These new compounds showed good benefit, but only in a subgroup of patients. The presence of KRAS mutations was demonstrated to be correlated with the absence of efficacy of anti-EGFR drugs. At the same time, a few studies also proposed to investigate EGFR gene copy number, BRAF and PIK3CA mutations, as well as PTEN protein expression, in order to increase the predictive power of the efficacy of EGFR-targeted therapies. Second aim of the present work was therefore to investigate in a cohort of SBA the aforementioned markers involved in the EGFR pathway, in order to verify if the pattern of these alterations could justify the possible introduction of these therapies also in patients affected by SBA.
All the proposed analyses will represent a significant improvement in the knowledge of the molecular characterization of SBA, which will lead to a better treatment of such a severe neoplastic disease.
PATIENTS AND METHODS
3.1 Patients
We recruited 40 patients affected by SBA. Due to the low frequency of occurrence of this type of neoplastic disease, we collected cases from different institutions.
- Sixteen consecutive patients were surgically resected in Canton Tessin and identified in the databases of the Institute of Pathology in Locarno, Switzerland, from 1996 to 2007.
- Five consecutive patients were identified in the Operative Unit of Pathology, Civil Hospital of Legnano (Milan, Italy), from 1997 to 2005.
- Twelve consecutive patients were identified at the Department of Medical Sciences, University of Eastern Piedmont “Amedeo Avogadro” of Novara (Italy), from 2001 to 2008.
- Seven consecutive patients were identified in the databases of Operative Unit of Pathology and Laboratory Medecine of Multimedica (Milan, Italy), from 2003 and 2009.
Patients with tumor of Papilla of Vater were excluded.
3.2 Patient treatment and clinical evaluation
A patient of our cohort was treated with irinotecan but at the end developed a metastatic lesion. Therefore a treatment with cetuximab was proposed.
The patient was treated at Oncology Institute of Southern Switzerland (Bellinzona, Switzerland) with cetuximab in combination with irinotecan with the following scheme:
- cetuximab: loading dose of 400 mg/m2 over 2 hours, followed by weekly 250 mg/m2 over 1 hour;
- irinotecan: same dose and schedule used at progression.
Treatment was continued until progressive disease (PD) or toxicity occurred, according to the standard criteria (Therasse et al, 2000) or to specific trial guidelines.
Clinical response was assessed every 6 to 8 weeks with radiologic examination (computed tomography or magnetic resonance imaging). The Response Evaluation
Criteria in Solid Tumors (RECIST) (Therasse et al, 2000) were adopted for clinic evaluation, and objective tumor response was classified as partial response (PR), stable disease (SD), or PD.
3.3 Molecular analyses
All the analyses were performed on tumor specimens fixed in 10% buffered formalin and embedded in paraffin (FFPE). FFPE tumor blocks were reviewed for quality and tumor content by analyzing detailed morphology of haematoxylin and eosin stained tissue sections of each blocks. A single representative tumour block from each case, containing at least 70% of neoplastic cells, was selected for immunohistochemical, cytogenetic and molecular analyses. Tumour macrodissection was performed in tumour blocks containing less than 70% of neoplastic cells (to reduce the presence of non-neoplastic tissues) following Van Krieken guidelines (Van Krieken et al, 2008).
3.4 Immunohistochemical analyses
Immunohistochemical analyses were performed on 3-μm thick tissue sections using a Benchmark automatic immunostaining device (Ventana Medical System, Tucson, AZ, USA). According to manufacturer’s instructions, specimens were incubated after heat induced antigen retrieval with the specific antibody. More in details, tissue sections were sequentially deparaffinizated with xylene and rehydrated in alcohol solutions and then in distilled water. Subsequently, tissue sections were incubated for 5 minutes with proteinase K, for 5 minutes with peroxidase block solution, for 30 minutes with primary antibody or negative control reagents and for 30 minutes with the secondary goat anti-mouse antibody and horseradish peroxidase molecule linked to a common dextran polymer. At the end, diaminobenzidine (DAB+) substrate chromogen solution was applied for 10 minutes and, after counterstaining with haematoxylin and coverslipping, the sections were kept in the dark at room temperature until the evaluation that was made by using a light microscope. Positive and negative controls were included in each slide run.
ββββ-catenin. β-catenin protein expression analysis was performed using anti-β-catenin monoclonal antibodies (BD Biosciences, San Jose, CA, USA) at 1:50 dilution as previously reported (Frattini et al, 2004). A normal β-catenin expression is considered
when tumoral tissues show a cytoplasmic staining similar to that observed in paired healthy mucosa. APC or β-catenin genes alterations lead to an abnormal accumulation of β-catenin protein in the cytoplasm, and a nuclear staining (due to the fact that when the β-catenin protein is too much accumulated in the cytoplasm is
able to enter into the nucleus) can also be observed. Therefore, we considered negative cases those showing a cytoplasmic expression of β-catenin protein superimposable with β-catenin expression of healthy mucosa. Then, patients with an
abnormally high expression of β-catenin protein were subdivided into 2 groups:
those showing overexpression only in the cytoplasm (and named Positive-cytoplasm or Pos Cyt), and those showing also expression in the nucleus (and named Positive- nucleus or Pos Nucl). Healthy tissue (i.e. normal small bowel mucosa) was used as internal control; colorectal adenocarcinoma with a nuclear expression of β-catenin was used as external positive control.
PTEN. PTEN protein expression analysis was performed using the anti-PTEN Ab-4 monoclonal antibodies (Neomarkers, Fremont, CA, USA) at 1:50 dilution as previously reported (Frattini et al, 2005; Saal et al, 2005; Frattini et al, 2007). PTEN staining intensity scores for invasive tumor and non-neoplastic cells were recorded as described in the literature (Saal et al, 2005) and on the basis of our experience (Frattini et al, 2007). PTEN protein expression was mainly detected at the cytoplasmic level, while very few cases also showed nuclear positivity. We considered PTEN negative tumours those showing a striong reduction or absence of immunostaining in at least 50% of cells, as compared with the internal control (i.e., vascular endothelial cells and nerves). Healthy tissue (i.e. normal colon mucosa) was used as internal positive control; normal endometrium was used as external positive control.
3.5 Fluorescent in situ hybridization (FISH)
EGFR gene status evaluation was performed on 3-μm thick tissue sections that were treated using Paraffin Pretreatment kit II (Abbott Molecular, AG Baar, Switzerland) according to manufacturer’s instructions. Dual-colour FISH assay was performed using LSI EGFR/CEP7 probes (Abbott Molecular). The LSI EGFR probe is labelled in SpectrumOrange and covers an approximately 300 kb region that contains the entire EGFR gene at 7p12 (red signals). The CEP7 probe, labelled in SpectrumGreen (green signals), hybridises to the alpha satellite DNA located at the centromere of Chromosome 7 (7p11.1– q11.1) (Figure 3.1).
a
7p12 LSI EGFR SpectrumOrang e
7p11.1-q11.1 CEP7 SpectrumGreen
b
Telomero Regione 7p12 Centromero
gene EGFR
a
7p12 LSI EGFR SpectrumOrang e
7p11.1-q11.1 CEP7 SpectrumGreen
b b
Telomero Regione 7p12 Centromero
gene EGFR
Figure 3.1. Visual representation of LSI EGFR/CEP7 dual colour probe; a: Probes position and types with respect to the entire Chromosome 7; b: Position of EGFR probe with respect to EGFR gene (www.abbottmolecular.com).
Target sections and probes were co-denatured at 75°C for 5 min and allowed to hybridize overnight at 37°C. A post-hybridisation stringency wash was carried out in a water bath at 72°C for 5 min. After washing twice and drying at room temperature for 10 min, slides were mounted with 406-diamidino-2-phenylindole (DAPI II; Vysis).
Fluorescent in situ hybridization signals were evaluated with a fluorescent automated microscope (Zeiss Axioplan 2 Imaging, Zeiss, Oberkochen, Germany) equipped with single and triple band pass filters. Image for documentation were captured using an AxioCam camera (Zeiss Axiocam MRm) and processed using the AxioVision system (Zeiss). Patients were classified using descriptive criteria, taking
into account the abnormalities revealed and the percentage of cells involved (Martin et al, 2009).
To overcome the problem of tissue heterogeneity, we evaluated 10 different tumour areas and at least 10 representative nuclei from each area. Overall, a total of 100 cells for each patient were scored. For cases in which only a biopsy was available, we evaluated all the analysable nuclei.
A normal cell is characterized by 2 red signals and 2 green signals (Figure 3.2a), or by 4 red signals and 2 green signals during DNA replication (Figure 3.2b).
Figure 3.2. Schematic representation of EGFR FISH on chromosomes and in cells. Green signals correspond to the centromere of Chromosome 7, red signals to EGFR gene. a: normal nucleus; b: normal nucleus with DNA replication.
In all the other situations there is an abnormal EGFR gene status, depending on different mechanisms:
1) alterations linked to the number of Chromosomes 7;
2) alterations of EGFR gene itself;
In the first category we can identify the following situations:
- loss of Chromosome 7 (also named monosomy) (Figure 3.3a);
- low polysomy (3 or 4 balanced copies of red and green signals) (Figure 3.3b);
- high polysomy (more than 4 balanced copies of red and green signals) (Figure 3.3c).
Figure 3.3. Schematic representation of EGFR alterations linked to the number of Chromosome 7). a: loss of Chromosome 7; b: low polysomy (for example, 3 copies of the Chromosome 7); c: high polysomy (more than 4 copies, in the example there are 6 copies, of the Chromosome 7).
EGFR alterations due to a problem of EGFR gene itself are caused by gene amplification. In general, in this situation, the number of red signals is higher than that of green signals, therefore the ratio (R) between red and green signals is more than 2. In this category, we can identify 3 different situations, although each of them leads to an abnormally high number of EGFR gene:
- low level of gene amplification (when 2<R<5) (Figure 3.4a);
- high level of amplification HSR-type (Homogenously-Stained Regions), when R>5 and red signals are in clusters (Figura 3.4b);
- high level of amplification DM-type (Double Minutes), when R>5 and red signals are dispersed in the nucleus and correspond to extra-chromosomal sequences without centromere (Figura 3.4c).
Figure 3.4. Schematic representation of EGFR gene amplificfation (R>2). a: low level of gene amplification (R<5); b: high level of amplification HSR-type (Homogenously- Stained Regions); c: high level of amplification DM-type (Double Minutes).
On the basis of criteria published for colorectal cancer (Martin et al, 2009; Varella Garcia et al, 2009), cases showing only 1 Chromosome 7 were classified as EGFR loss. When we observed 2 Chromosomes 7 in more than 60% of cells, the tumor was classified as disomic. Tumour samples with an aberrant number of Chromosomes 7, defined as ≥ 3 copies in at least 40% of cells, were classified as polysomic and: low polysomic in the presence of 3 copies of Chromosomes 7, tetrasomic in the presence of 4 copies of Chromosomes 7, highly polysomic in the presence of more than 4 copies of Chromosomes 7. Specimens with a R>2 between EGFR gene and Chromosomes 7 centromere signals in at least 10% of cells were classified as carrying EGFR gene amplification. As patients carrying either at least a tetrasomic profile or gene amplification show a significant gain of EGFR gene, we grouped them into a class named copy number gain (CNG), according to the criteria described in the literature with slight modifications (Moroni et al, 2005): this group is also considered to be FISH positive (FISH+). Loss of Chromosomes 7, disomy as well as low polysomy are cumulatively considered as FISH negative (FISH-). Table 3.1 includes all classifications we applied.
Categories Anomaly Number of cells Status
Loss (L) 1 copy of Chromosome 7 > 50% FISH-
Disomy (D) 2 copies of Chromosome 7 > 60% FISH-
Low Polysomy (LP) 3 or 4 copies of Chromosome 7 > 40% FISH- FISH+(*) High polysomy (HP) more than 4 copies of Chromosome 7 > 40% FISH+
Gene Amplification (A) R>2 > 10% FISH+
Table 3.1. Interpretation criteria for EGFR FISH (Martin et al, 2009; Varella Garcia et al, 2009).
Legend: R: ratio between gene and centromere signals (red and green signals, respectively), (*) only if tetrasomy.
3.6 Mutational analysis by direct sequencing
Genomic DNA was extracted using the QIAamp Mini kit (Qiagen, Chatsworth, CA, USA) according to manufacturer’s instructions.
KRAS (exons 2 and 3), BRAF (exon 15), PIK3CA (exons 9 and 20) and TP53 (exons 4- 10) gene mutations were detected by direct sequencing on genomic DNA as already reported (Frattini et al, 2004; Frattini et al, 2005; Di Nicolantonio et al, 2008). KRAS exon 2 includes codons 12 and 13, KRAS exon 3 includes codon 61, BRAF exon 15 includes codon 600, PIK3CA exon 9 includes codons 542 and 545 and PIK3CA exon 20 includes codon 1047. All these codons represent sites where the large majority of oncogenic mutations occur (Davies et al, 2002; Frattini et al, 2004; Samuels et al, 2004). As for TP53 gene, there are not specific hotspots (although a higher percentage of alterations is observed in codons 175, 245, 248 and 273), and the mutations can occur in all the nucleotides between exon 4 and exon 10: therefore we amplified all these exons. The nucleotide sequence corresponding to every exon was amplified from tumour-extracted genomic DNA by Polimerase Chain Reaction (PCR), purified (Microcon YM-50, Millipore, Billerica, MA, USA) and directly sequenced. The list of primers used for mutational analyses is reported in Table 3.2.
All samples were subjected to automated sequencing by ABI PRISM 3130 (Applied Biosystems, Foster City, CA, USA). All mutated cases were confirmed at least twice
starting from independent PCR reactions. In each case, the detected mutation was confirmed in the sequence as sense and antisense strands.
Gene Exon Forward Primer Reverse Primer Annealing
Temperature (°C) K-Ras 2 5’-TGGTGGAGTATTTGATAGTGTA-3’ 5’-CATGAAAATGGTCAGAGAA-3’ 55
K-Ras 3 5’-GGTGCACTGTAATAATCCAGA-3’ 5’-TGATTTAGTATTATTTATGGC-3’ 49 BRAF 15 5’-TCATAATGCTTGCTCTGATAGGA-3’ 5’-GGCCAAAAATTTAATCAGTGGA-3’ 52 PIK3CA 9 5’-GGGAAAAATATGACAAAGAAAGC-3’ 5’-CTGAGATCAGCCAAATTCAGTT-3’ 56 PIK3CA 20 5’-CTCAATGATGCTTGGCTCTG-3’ 5’-TGGAATCCAGAGTGAGCTTTC-3’ 55 TP53 4.1 5’-GAGGACCTGGTCCTCTGACT-3’ 5’-AAGGGACAGAAGATGACAGG-3’ 60 TP53 4.2 5’-AGAGGCTGCTCCCCGCGTGG-3’ 5’-ATACGGCCAGGCATTGAAGT-3’ 60 TP53 5 5’-TTCAACTCTGTCTCCTTCCT-3’ 5’-CAGCCCTGTCGTCTCTCCAG-3’ 62 TP53 6 5’-GCCTCTGATTCCTCACTGAT-3’ 5’-TTAACCCCTCCTCCCAGAGA-3’ 62 TP53 7 5’-AGGCGCACTGGCCTCATCTT-3’ 5’-TGTGCAGGGTGGCAAGTGGC-3’ 64 TP53 8 5’-TTCCTTACTGCCTCTTTGCTT-3’ 5’-AAGTGAATCTGAGGCATAAC-3’ 56 TP53 9 5’-AGCAAGCAGGACAAGAAGCG-3’ 5’-ACTTGATAAGAGGTCCCAAG-3’ 58 TP53 10 5’-TTTTAACTCAGGTACTGTGT-3’ 5’-CTTTCCAACCTAGGAAGGCA-3’ 58
Table 3.2. Genes analyzed, primers sequences and annealing temperatures.
3.7 MSI
The status of MSI was assessed by the analysis of the microsatellite loci included in the panel of Bethesda (BAT25, BAT26, D2S123, D5S346, and D17S250), as reported in the literature (Frattini et al, 2004). The list of primers used for MSI analysis is reported in Table 3.3. MSI was confirmed by the presence of additional peak(s) in tumor sample compared with the pattern of the normal paired tissue. MSI was defined as being present when more than 30% of investigated loci showed instability.
Exon Forward Primer Reverse Primer Annealing
Temperature (°C) BAT25 5’-TCGCCTCCAAGAATGTAAGT-3’ 5’-TCTGCATTTTAACTATGGCTC-3’ 52
BAT26 5’-TGACTACTTTTGACTTCAGCC-3’ 5’-AACCATTCAACATTTTTAACCC-3’ 52 D2S123 5’-AAACAGGATGCCTGCCTTTA-3’ 5’-GGACTTTCCACCTATGGGAC-3’ 52 D5S346 5’-ACTCACTCTAGTGATAAATCGGG-3’ 5’-AGCAGATAAGACAGTATTACTAGTT-3’ 52 D17S250 5’-GGAAGAATCAAATAGACAAT-3’ 5’-GCTGGCCATATATATATTTAAACC-3’ 52
Table 3.3. Loci analyzed, primers sequences and annealing temperatures. Forward primers are labeled with 6-FAM at 5’-end
3.8 Allelic imbalance of Chromosome 18q
Several tumor suppressor genes are located on Chromosome 18. Of these, 3 genes play a relevant role in CRC: DCC, SMAD2 (previously named JV18 or MADR2) e SMAD4 (previously named DPC4 or MADR4), located on the long arm of Chromosome 18 (i.e: q portion of the chromosome). To evaluate if Chromosome 18q region is lost, thus meaning that these tumor suppressor genes are absent and therefore cannot exert their biologic activity, it is useful to investigate the allelic imbalance, also named as loss-of-heterozygosity.
On the basis of our previous experience (Frattini et al, 2004), we included in our analysis the following loci: D18S64, D18S484, D18S474, D18S1110, D18S1161, D18S68, D18S1102 (see Figure 3.5 for the location of these loci with respect to DCC, SMAD2 and SMAD4 genes).
D18S1110
D18S474 D18S64 D18S1161 D18S1102
D18S68
CENTROMERO
DCC SMAD4
SMAD2
D18S484 D18S1110
D18S474 D18S64 D18S1161 D18S1102
D18S68
CENTROMERO
DCC SMAD4
SMAD2
D18S484
Figure 3.5. Loci position with respect to the centromere of Chromosome 18q and with respect to DCC, SMAD2 and SMAD4 genes.
These loci were analyzed according to the protocol already published (Frattini et al, 2004). The list of primers used for allelic imbalance of Chromosome 18q is reported in Table 3.4.
Exon Forward Primer Reverse Primer Annealing Temperature (°C) D18S64 5’-ATACTGGTGGTGGTTATACAACAT-3’ 5’-AAATCAGGAAATCGGCA-3’ 52
D18S484 5’-TGTAGCATTTTTAAGACAGTAAAG-3’ 5’-ACATATTCCTTGCTTTGTCA-3’ 52 D18S474 5’-TGGGGTGTTTACCAGCATC-3’ 5’-TGGCTTTCAATGTCAGAAGG-3’ 52 D18S1110 5’-TGACCTTGGCTACCTTGC-3’ 5’-TCGAAAGCCTTAAACTCTGA-3’ 52 D18S1161 5’-GTCCGTCCAACGTCCAA-3’ 5’-GGAGAGCCACACCTATCCTG-3’ 52 D18S68 5’-ATGGGAGACGTAATACACCC-3’ 5’-ATGCTGCTGGTCTGAGG-3’ 52 D18S1102 5’-TTTCAGGATTTTGGAGCC-3’ 5’-GGAATGACTGCGTCTGTG-3’ 52
Table 3.4. Loci analyzed, primers sequences and annealing temperatures. Forward primers are labeled with 6-FAM at 5’-end.
The analysis requires the availability of healthy tissue. If in the healthy tissue sample the 2 alleles have the same size, the patient is homozygous for that locus and cannot be evaluated for the analysis of allelic imbalance (the locus is classified as “not informative” (NI)).
If in the tumoral tissue there is the presence of additional peak(s) compared with the normal paired tissue, the locus is considered “instable” (or MSI), and again it is considered NI for the analysis of allelic imbalance.
If in both normal and tumoral tissue there are 2 alleles with different sizes, the patient is heterozygous for that locus and the allelic imbalance can be ascertained, using the following formula:
N1 T2 R = --- x ---
N2 T1
where
N1 = height of lower dimension peak of healthy tissue N2 = height of higher dimension peak of healthy tissue T1 = height of lower dimension peak of tumoral tissue T2 = height of higher dimension peak of tumoral tissue.
As cut-off, on the basis of the literature (Frattini et al, 2004), we used the value of 30%. Therefore, if R is between 0.7 and 1.3, the locus is normal, not lost. If R ≤ 0.70 or
R ≥ 1.3, there is allelic imbalance, that means that 1 allele is lost. A patient is classified as carrying the loss of the Chromosome 18q region when at least 30% of informative loci have R ≤ 0.70 o R ≥ 1.3.
3.9 Statistical analyses
For all the statistical correlations, we used the Fisher’s exact test because it is appropriate for the analysis of small number of cases and for zero values. The level of significance was set at p=0.05.
